Towards a framework for Requirements Engineering in agile Global Software Development

Table of Contents

List of figures, tables and abbreviations

1. Introduction

2. Objectives and methodology

3. Overcoming Requirements Engineering and Management barriers in GSD

3.1. Global Software Development

3.2. Agile methods

3.3. Requirements Engineering and Management

3.4. Cultures

4. Requirements Engineering Framework

5. Application

6. Discussion

7. Conclusion/Outlook

References

Appendix

List of figures, tables and abbreviations

Figure 1: Scrum process, source: (Cohn, 2013)

Figure 2: Comparison of five cultural dimensions for Germany, India and the United States, Source: (Hofstede, 2013b)

Table 1: GSD collaboration barrier, created distance types and barrier impact on collaboration performance

Table 2: Adapted from: (Damian, 2007). Classification of multi-site settings

Table 3: Mapping of GSD problems, RE-challenges and affected RE activities (Damian et al. 2003)

Table 4: RE challenges and success factors

Table 5: The GLOBE scores for Germany, India and the United States

Table 6: RE/RM framework.

Table 7: Requirements Engineering Framework; mapping interventions to RE activities and GSD barriers

Table 8: Barriers for BizAppMaker case.

Table 9: Interventions mapped to RE/RM activities and barriers

Abbreviations

illustration not visible in this excerpt

1. Introduction

According to Gartner (Pettey, 2012) the worldwide IT-outsourcing services spending surpassed $251 Billion in 2012. As software is of such great importance for the success of businesses, many organizations began to outsource development or started to create remotely located – so called offshore - software development subsidiaries. Reasons for this trend are according to Herbsleb and Moitra (Herbsleb and Moitra, 2001: 16/17) that businesses seek proximity to the growing markets in order to benefit from the knowledge about customers and local conditions. Moreover, the pressure to realize shorter time-to-market makes companies develop in several shifts 24 hours per day around the globe. Furthermore, companies want to benefit from a global resource pool (especially if there are skilled labor forces in low-wage countries).

But there is a high risk of a failure for endeavors of companies that conduct global software development (GSD) – especially when the Global distance (geographical, temporal, linguistic or cultural distance) of customers, vendors or development teams is high. Risks and challenges are multiplied in the GSD context. For example the risk of a communication breakdown or lack of trust that leads to low efficiency because of rework or later delivery of the project.

With a market share of almost 53% the United States is the biggest software market (IDC, 2012). India is a huge outsourcing and offshoring destination (DiamondCluster International, 2005). Often European companies play a central role in a networked business model and coordinate a software project interfacing American customer and Indian developer. Therefore we have chosen these countries for our research. The cultural gap between America and India is big. Carmel and Abbott (2007) call India “distant lands” due to the difficulties with distance management and cultural differences (Carmel and Abbott, 2007: 45).

Agile methodologies have increased in popularity during the last years (Hillegersberg et al., 2011: 134), (Abrahamsson et al., 2002), (Larsen and Shore, 2012). A growing number of companies also use agile methods in global software development settings (Ramesh et al., 2006). Agile methods rely heavily on communication instead of documentation and therefore seem to contradict with the nature of GSD e.g. dispersed teams, little overlap of working hours for synchronous meetings and lack of informal communication.

One, if not the most critical, phase in a software project is the requirements engineering (RE). According to a research conducted by the Standish Group 13.1% of projects fail due to incomplete requirements. Further reasons for project failure which have a direct relation to RE are lack of user involvement, unrealistic expectation and changing requirements/specifications (Hull et al., 2011: 3). These figures underline the importance of RE and requirements management (RM) activities in software projects. RE is a very communication and collaboration intense area, which seems to oppose to the characteristics of GSD e.g. geographical and time difference, which hinder synchronous communication, negotiation and collaboration.

Based on a literature review this paper analyses barriers impeding typical RE and RM activities in agile GSD settings. The specific RE and RM-related challenges for the context of US customer and Indian developer in an agile software development setting are analyzed and a framework of methods and routines as interventions for overcoming the identified RE- and RM-barriers are proposed. The chosen domain is business user, i.e. business software developed for organizations.

A framework to overcome RE- and RM-challenges in agile GSD-settings could help both practitioners and academia. The RE-practitioners working in agile GSD projects can profit from a “toolbox” of interventions for the challenges they typically face in their daily assignments. Thus, the toolbox also helps organization to decrease the costs in case things go wrong, which arise out of longer project timelines, rework or even failure of projects. For academia this framework contributes to the current discourse in the field of requirements engineering and management in agile global software development.

2. Objectives and methodology

The addressed problem is of practical nature, which occurs in companies that apply GSD, e.g. due to cultural differences, geographical distance and existing barriers for collaboration. The desired output of the paper should be an artifact in the form of a framework that can be applied in practice to overcome challenges in GSD. When starting an endeavor like GSD, employees have a need for training and leading guidelines. Here, a framework could support the process, when starting an offshore project.

The research methodology used here is Design Science Research (DSR) (Peffers et al., 2007: 12ff.), which is applied to an illustrative scenario (Peffers et al., 2012: 402). Design science aims at the creation and evaluation of IT artifacts in order to solve organizational problems (March and Storey Veda, 2008: 727). An illustrative scenario is the application of an artifact to synthetic situation or a real-world situation with the objective to prove that the artifact is suitable (Peffers et al., 2012: 402).

There are two reasons, why we have chosen DSR as our research method. First, as Pfeffers (2007) points out that DSR is important in a discipline that is oriented towards the successful creation of artifacts (Peffers et al., 2007: 48) and the outcome of this paper is a framework, DSR fits the objective very well.

Second, Pfeffers (2007) calls design “the act of creating an explicitly applicable solution to a problem” (Peffers et al., 2007: 49). This shows that the focus of DSR is the creation of an artifact that can be applied to achieve a solution. Again, the chosen method fits the objective of the paper to create a framework, which can be applied by GSD practitioners in their real-world assignments.

There is hitherto no comprehensive framework for overcoming RE challenges in this specific setting that develops interventions based on cultural models, GSD- and agile barriers, which is why we aim at filling the gap here.

DSR activities include:

1. Problem identification and motivation
2. Definition of objectives for a solution
3. Design and development
4. Demonstration
5. Evaluation
6. Communication

Following the problem identification in the first chapter the objective of the solution is defined in this chapter. It is to create a framework which proposes interventions for specific culture related issues when conducting RE/RM activities in agile-style GSD projects. The framework will contribute to a qualitative improvement of the "toolkit" available for requirements engineering practitioners and scientists working in multicultural, agile GSD settings.

The research questions for this paper are:

- Which are the main GSD barriers in agile GSD settings?
- Which are the typical requirements engineering activities these barriers hinder?
- Which are the challenges stemming from cultural differences within collaboration of members of the Indian, German and US society?
- Which interventions can be applied to overcome barriers when elaborating and managing requirements in the context of agile-driven Global Software Development in a setting when interfacing American users and Indian developers?

The design and development activity starts in chapter three with capturing of theoretical knowledge about the barriers for GSD, challenges of agile methods in GSD, cultural challenges within the interaction as an interface to American user and India developers and typical RE activities. Then in chapter 5 we construct the framework using a design approach to identify the relation between GSD barriers, RE activities as well as cultural challenges and propose interventions to overcome the identified challenges.

The demonstration activity is applied in chapter five. In order to show the suitability of the framework for the above described issues, it is applied in a synthetic situation. It is demonstrated how the proposed interventions address RE challenges. Due to the limited time and capacity for this project paper it is not possible to apply the proposed solution in a real world experiment.

Following a comparison of the objectives to the observed results of the illustrative scenario the proposed framework is critically discussed and the need for further research is identified and described in the outlook.

Also due to the limited time the evaluation and communication activities of the DSR method are not covered by this paper, and thus lead to need for further research.

In this chapter we defined the objectives for the intended solution. In the next chapter we are creating the theoretical basis for the framework.

3. Overcoming Requirements Engineering and Management barriers in GSD

3.1. Global Software Development

Global Software Development (GSD) can be defined “as software development that uses teams from multiple geographic locations” (Sangwan et al., 2007: 3). Damian (2007) synonymously uses the term Global Software Engineering (GSE) (Damian, 2007: 21). Herbsleb states that the key phenomenon of GSD is coordination over distance (Herbsleb, 2007: 188).

According to Carmel (1999) motivators for GSD are:

- Limited specialized talents for needed technologies,
- A trend of global mergers and acquisitions in the software industry,
- Geographical differences in development cost,
- A development work system based on shifts enabling shorter time to market
- Proximity to local markets (Carmel, 1999: 4ff.).

In the literature different terms are used for teams in the GSD context. Carmel (1999) states “a Global Software Team is separated by a national boundary while actively collaborating on a common software/systems project” (Carmel, 1999: 3). Dafoulas (2002) depicts a more detailed picture of virtual teams: they are often distributed and commonly resources are geographically dispersed, which means that they are operating in different time zones. Team members often belong to different organizations and thus companies, but still work together in one project. The organization of virtual teams is a network structure. They need frequent as well as structured communication, for which they predominantly use electronic means. In order to operative effectively virtual teams need a high level of trust (Dafoulas and Macaulay, 2002.: 2). According to Dafoulas the difference between the various terms is very fuzzy. We would therefore like to use the term global team to be in line with global software development.

Challenges, barriers and success factors

When conducting GSD projects organizations face various challenges and barriers that hinder project success. According to the Oxford dictionary a challenge is a task that test’s someone’s abilities (2013). We therefore define a GSD challenge as an organizational task or endeavor in a GSD setting that is testing the abilities of the organization and its members. Pirkkalainen and Pawlowski (2012) define a barrier as: “any challenge, risk, difficulty, obstacle, restriction or hindrance that might prevent a single person, a group or an organization to reach an objective and success in a specific context when the challenge is related to acting or working in a collaborative cross border setting” (Pirkkalainen and Pawlowski J., 2013: 5). A GSD success factor is a factor that needs to be taken into account or fulfilled in order to mitigate overcome barriers and succeed in a GSD challenge.

A GSD challenge Carmel (1999) states as one of the centrifugal forces of global teams is dispersion. He argues that it is better to manage co-located teams rather dispersed teams. Co-location allows managers to manage by walking around, whereas this in not possible in dispersed settings. This implies the risk of “out of sight is out of mind”. Moreover, costs are higher and probability of communication is lower, if teams are dispersed (Carmel, 1999: 42).

Herbsleb (2007) states that the essential difference of GSD settings to co-located settings is that most of the mechanisms of coordination are absent or not working due to fact that the teams are dispersed (Herbsleb, 2007: 189). Coordination is defined as the act of managing dependencies between activities (Malone and Crowston, 1994: 90). Carmel (1999) argues that coordination breakdown is one of the five centrifugal forces of global teams. Due to dispersed teams the informal coordination mechanisms are hindered, for instance it is not possible to simply approach a colleague in the office. Moreover, due to the rise of coordination needs for global teams, the need for communication rises, too (Carmel, 1999). In a GSD setting finding the right person to solve an issue is a time-consuming endeavor, which often leads to making assumptions in the meanwhile. These assumptions tend to be in conflict with assumptions or requirements on other localities of the distributed project (Sangwan et al., 2007: 5).

Further challenges are late response or low urgency to inquiries from unknown personnel, unavailability because of local holidays, differing level of motivation between teams, fear of becoming redundant or losing their job after completion of the project (Sangwan et al., 2007: 5).

Herbsleb (2007) also mentions much less communication and less effective communication as GSD challenges e.g. due to little overlapping hours, replies only arrive on the next day (Herbsleb, 2007: 189). Moreover, there is the challenge of inadequate communication. Either the formal, official communication can suffer from a poorly specified interface, which leads to loss of time and not-identified problems. Or the informal communication, for instance the spontaneous talk in the corridor, which usually helps people to stay aware of what other people are engaged with, is much less in GSD settings (Herbsleb and Moitra, 2001: 18). Carmel (1999) states that the loss of communication richness is a centrifugal force for global teams. People want to use richer communication for certain tasks, e.g. face-to-face is used to solve a difficult dispute. As 80% of the communication in non-verbal we like to use richer communication to communicate more information. This is especially important for high context (HC) cultures but also for the other cultures (Carmel, 1999: 48).

Language and cultural aspects are a source of challenges, too. For example, non-native English speakers prefer addressing concerns via email instead of participating actively in telephone conferences. Persons from a direct culture are often perceived rude by persons from an indirect culture, when directly addressing an issue via phone or email, while persons from direct cultures find it frustrating, when people from an indirect culture make the impression of not getting to the point (Sangwan et al., 2007: 5–6). Often serious and constant misunderstandings arise out of cultural issues due to different communication style. Cultural diversity often makes communication even more difficult, when people are irritated and don’t know how to respond (Herbsleb and Moitra, 2001: 17).

Further challenges comprise time consuming logistical issues, aligning management practices, organizational processes and setting up technical infrastructure (Sangwan et al., 2007: 6). Different sites in GSD projects often use different development tools, processes, and practices and have different habits. This leads to various problems including confusion, misunderstanding, unclear status and inconsistent information (Herbsleb, 2007: 189).

In GSD settings participants suffer from a lack of awareness , because people from different sites know little about what people at other remote sites are working on. This leads to misunderstandings and hinders the ability to keep track of changes (Herbsleb, 2007: 189).

Another challenge for GSD settings with dispersed teams is loss of “teamness ”. Carmel (1999) argues that global teams suffer a loss of cohesion as there is less trust, more stereotyping and more in-group conversation. Moreover, due to the risk of miscommunication global teams are less relaxed with each other, as they permanently need to be aware of linguistic and cultural messages (Carmel, 1999: 53).

Herbsleb and Moitra (2001) argue that knowledge management is a challenge due to physical distance in GSD settings. If new information from customers or the market is not shared directly within the project, then teams cannot adapt their priorities accordingly. Poor documentation tends to cause inefficiencies in the collaborative development as it leads to assumptions and ambiguity (Herbsleb and Moitra, 2001: 18).

Project and process management issues are further challenges for GSD. For instance a lack of synchronization between remote sites can lead to critical situations in the software test. Moreover, volatile requirements and unstable specifications lead to difficulty in implementing the same engineering principles across sites (Herbsleb and Moitra, 2001: 18).

Prikladnicki (2006) states that requirements engineering (RE) is the main challenge in the software development process as the distance between the teams negatively affects the understanding and agreement (Prikladnicki et al., 2006: 22f).

As defined above are barriers any challenges that hinder people and organization to reach an objective in a cross border-collaboration situation. Different barrier frameworks can be found in the literature, which differ regarding their context and focus. In the following we describe and discuss two barrier frameworks from Noll et al. (2010) and Pallot (2010). Noll’s framework covers the context of virtual teams and collaboration in GSD projects and covers the GSD literature (Noll et al., 2010). Pallot’s framework is intended for the context of proximity and distance in online collaboration between professionals and has a more holistic and detailed view, because it considers a broad range of research areas like Community of Practice, Computer Support for Cooperative Work, Distributed Knowledge Management, Geographic Dispersion in Teams, Front-End Innovation, Inter-Organizational Collaboration, Knowledge Management, and New Product Development (Pallot et al., 2010b: 1–6).

Based on a systematic literature review Noll et al. (2010) identified eight barrier categories for collaboration in GSD:

- Language and Cultural distance
- Temporal distance
- Geographic distance
- Process and management issues
- Fear and trust
- Infrastructure
- Organization
- Product architecture (Noll et al., 2010: 72).

Based on a multi-disciplinary literature review Pallot introduces four barrier dimensions (structural, social, technical and legal) containing 20 different distance types. Distance is created by barriers. The following table shows a mapping of GSD barriers from both frameworks (Pallot 2010, Noll et al. 2010) to created distances and the barrier impact based on a survey conducted by Pallot (Pallot et al., 2010a):

Table 1: GSD collaboration barrier, created distance types and barrier impact on collaboration performance

illustration not visible in this excerpt

Based on Pallot et al. (Pallot et al., 2010b) comprehensive 18 distance type barrier framework the following section presents distance types for GSD. The barrier framework comprises four barrier dimensions namely structural, social, technical, legal & ethical dimension.

The structural dimension includes four different distance types i.e. configurational, institutional, organizational, spatial and temporal.

Configurational distance deals with the arrangement of group members across sites. A fully dispersed configuration would be, if only one team member is placed in one site. Sub-group configurations can lead to conflict, if members are left out of group communication and interactions. Distance-creating factors are dispersed teams, lack of leadership or incentive, lack of shared vision (Pallot et al., 2010b: 7). A lack of leadership hinders collaboration performance, because it is not clear who is leading the leading the collaboration process. This cause a lack of shared vision, goals and objectives of the collaborating partners (Pallot, 2009). Karolak (1998) states that lack of leadership leads to a situation in which remote teams fulfill their own interests (Karolak, 1998: 119). Distance bridging factors for configurational distance are clear leadership, shared vision, collaboration incentive, balanced power and expertise in decision making (Pallot et al., 2010b: 7).

Pallot (2009) argues that geographical dispersion of team members creates institutional distance is (Pallot, 2009). Institutional distance between group members is created through their belonging to different historical, political, economic and cultural or social setting. This affiliation brings about certain formal rules applied by its members. Distance-creating factors are globalization, different legal framework or political climate and lack of interoperability. Distance bridging factors for institutional distance are international experience or professional training (Pallot et al., 2010b: 7).

Pallot (2009) states that a multi-disciplinary setting creates organizational distance. In a multi-disciplinary setting different participants are involved, which creates diversity leading to a management overhead and additional effort (Pallot, 2009). Organizational distance is the degree to which rules and behavior that improve coordination activities are different. If members don’t belong to the same structure and don’t use the same routines they are organizationally distant. Distance-creating factors are to not belong to the same group and lack of behavioral cohesion (Pallot et al., 2010b: 7). Noll et al. (2010) argues that organization barriers in GSD create challenges beyond distance and cultural differences (Noll et al., 2010: 73). For instance, Bhat et al. (2006) point out that that in GSD projects the project-execution power, the decision-making power and the group with the knowledge of the stakeholder needs are often dispersed (Bhat et al., 2006: 38). Distance-bridging factors for organizational distance are to belong to the same cluster for example of professionals and multidisciplinary communities (Pallot et al., 2010b: 7).

Noll et al. (2010) identified 12 papers mentioning geographic distance as a barrier for collaboration (Noll et al., 2010: 72). Geographical dispersion of team members creates spatial distance (Pallot, 2009). Spacial distance makes collocation impossible. Closeness supports a shared vision, knowledge sharing and understanding, while remote working is said to be a barrier for shared vision and understanding as well as knowledge sharing (Pallot et al., 2010b: 7). Spatial and geographical distances are synonyms and in the further according to Pallot et al. (2010) we prefer to use the term spatial. Spatial distance creates a lack of informal meetings and therefore less opportunity for knowledge exchange and the development of personal relationship (Noll et al., 2010: 71), (Bhat et al., 2006: 41), (Carmel and Agarwal, 2001: 23–24), which hinders exchange of implicit knowledge (Noll et al., 2010: 71).

Pallot (2009) states that team members that work in different time zones as well as shift working creates temporal distance (Pallot, 2009). Further distance-creating factors are the lack of collocation and lack of face-to-face communication (Pallot et al., 2010b: 8). Noll et al. (2010) argues that temporal distance creates a barrier to collaboration in GSD as time zone difference limits the amount of synchronous meetings for development teams in different locations around the world. Another implication of temporal distance is the delay in response to asynchronous communication. Inquiries from one site to another, which arrive after working hours, can only be answered on the next day (Noll et al., 2010: 72). Distance-bridging factors are working in asynchronous mode. Distance-compression factors for temporal distance are the use of collaboration tools (Pallot et al., 2010b: 8).

From June 2007 to June 2008 a survey on collaboration barriers was conducted by ECOSPACE on the Ami Communities website accessible for 2000 registered members – mostly experienced researchers, developers and engineers - of the AMI@Work community. Regarding the structural dimension an amount of the 82 participants answered that a lack of clear leadership and a lack of incentive have the most significant impact on collaboration performance. Interestingly geographical dispersion and collaborating over different time zones were seen as less significantly affecting collaboration performance (Pallot et al., 2010a).

The social dimension comprises five different distance type namely relational, cultural, emotional, lingual and cognitive.

The barrier category “fear and trust” from Noll et al. (2010) contains factors that create relational distance. Piri et al. (2009) argues that geographic, temporal and cultural distance has a negative impact on trust as they limit the opportunity for informal face-to-face contact. This leads to fewer opportunities for building personal relationships, which support trust among people (Piri et al., 2009: 110). A lack of trust creates relational distance (Pallot, 2009). Relational distance is about how people build relationships with one another (Pallot et al., 2010b: 16). The human, intellectual and social capital enables different levels of relationship between individuals. If the relational distance is great, then the level of mutual trust is low. Mutual trust is needed to enable knowledge sharing and creation. Distance-creating factors are lack of social interaction and trust. Distance-bridging factors are trusted relationships and perceived similarity. Distance compression factors are social translucence and social awareness (Pallot et al., 2010b: 8). Socially translucent systems are systems that support visibility, awareness and accountability, so that people can have a social experience and interact with each other (Erickson and Kellogg, 2000: 59).

Pallot (2009) states that a lack of common usage and norms due to different geographical areas, business or organization cultures, creates cultural distance (Pallot, 2009). Noll et al. (2010) state that cultural distance is a barrier for GSD (Noll et al., 2010: 72). Cultural distance is created by different norms and values of individuals from different origin. It influences the behavior of individuals as well as their thoughts and interpretation. Distance-bridging factors are experience in international projects and education abroad (Pallot et al., 2010b: 9). Homogenous groups need more time to achieve a common understanding then heterogeneous groups, but the creativity stimuli in heterogeneous groups is greater than in homogenous ones (Pallot et al., 2010b: 17). For instance, expressions of acknowledgement of Asians might be interpreted as commitment from European and American people (Herbsleb et al., 2005: 531). Herbsleb (2005) argues that culture affects the interpretation of requirements as the domain knowledge, used to provide context for the requirements, is very different depending on the national culture (Herbsleb et al., 2005: 529).

Pallot (2009) argues that emotional behavior creates emotional distance (Pallot, 2009). Emotional distance is the degree to which individuals or groups recognize each other’s emotional states (Pallot et al., 2010b: 17). Emotional distance negatively affects the willingness of individuals to voice opinions and suggestions. More distance creates objectivity; less distance creates sympathy for collocated individuals. Distance-creating factors are emotionality; distance-bridging factors are cognitive training (Pallot et al., 2010b: 9). Damian argues that emotional distance supports requirements negotiation as there is a certain risk in face-to-face situations that emotions are negatively affecting the successful negotiation of requirements (Damian, 2001).

Pallot (2009) points out that multi-lingual setting creates lingual distance (Pallot, 2009). Lingual distance affects the possibility to share meanings and understanding within a group of heterogeneous individuals but also adds diversity as language influences behavior. Distance-creating factors are different languages; distance-bridging factors are a shared language (Pallot et al., 2010b: 9). Also, Noll et al. (2010) argues that language is a barrier for collaboration in GSD. Team members with less English language skills may prefer writing text messages like instant messaging or email over having telephone conferences or video conferences, which results in a loss of richness e.g. regarding visual or auditory signals. Poor language skills might be misinterpreted as lack of competence resulting in a less powerful situation and unwanted intimidation (Noll et al., 2010: 72).

Lack of media naturalness creates cognitive distance. When collaborating global team members can’t use all their natural senses to communicate, mimic and gesture get lost, creating a lack of media naturalness occurs (Pallot, 2009).

The ECOSPACE survey results for the social dimension show that lack of mutual trust is the most significant factor affecting collaboration performance. Though, other factors like interpersonal awareness also play a significant role for collaborating effectiveness as it is important to know what other team members are currently working on in order to adjust one’s own work accordingly. Interestingly the survey participants regarded emotional behavior as less significant for the collaboration performance (Pallot et al., 2010a). As stated above Damian found out that emotional distance in requirements negotiation situations avoids certain risks usually impeding negotiation success in face-to-face settings (Damian and Zowghi, 2002).

The technical dimension comprises five different distance type namely conceptual, contextual, referential, semantic and technological.

Pallot (2009) argues that a lack of common description e.g. a common concept, references or taxonomy creates conceptual distance (Pallot, 2009). Conceptual distance is the degree to which individuals associate the same meaning behind a certain concept. Conceptual distance within one community of practice makes is necessary to build a common ground and a shared understanding. Distance-creating factors are expertise gaps such as novice vs. expert. Distance-bridging factors are to create a common foundation and a shared meaning, for example by distance-compressing factors such as topic maps within wikis (Pallot et al., 2010b: 10).

Pallot (2009) states that a lack of operational mode e.g. context awareness creates contextual distance (Pallot, 2009). Contextual distance is affecting problem resolution, because of a lack of the right information relevant for the specific information (Pallot et al., 2010b: 18). Distance-creating factors are different local situations with differing circumstances and regulations resulting in a cognitive overload. Distance-bridging factors are context awareness, which can be created by accumulation of relevant information for example by deduction from shared events and meta-data (Pallot et al., 2010b: 10).

Pallot (2009) argues that multi-platform setting for example different types of devices in use or different operating systems create referential distance and technological distance (Pallot, 2009). Referential distance is depending on references in documents to the point of origin. The potential relevance of a document can be measured by the number of references. The higher the referential distance the lower is the relevance of the document. Distance-creating factors are the lack of information about the relevance of a document. Distance-bridging factors are the formulation of the correlation of the document e.g. by a program that calculated the relevance of documents (Pallot et al., 2010b: 10).

Pallot (2009) argues the unbalanced use of different technologies like Information and Communication Technology (ICT) or a different level of technological expertise creates technological distance (Pallot, 2009). Distance-creating factors are differing technological skills. Noll et al. (2010) point out infrastructure as a barrier for collaboration in GSD. In GSD settings global teams heavily rely on video and teleconferencing as a substitute for the lack of face-to-face contact (Noll et al., 2010: 73). Taweel at al. (2009) state that meetings simply cannot take place, when there is a failure of the equipment, which leads to delays in the project (Taweel et al.: 373). Noll et al. (2009) points out that due to geographic and temporal distance the implicit knowledge in the global team must be made explicit (Noll et al., 2010: 73). If the knowledge management infrastructure is not sufficient, it leads to a lack of shared understanding among dispersed teams (Damian and Zowghi, 2002: 323). A further barrier that creates technological distance is product architecture (Noll et al., 2010: 73). Herbsleb et al. (2003) state that the allocation of components to teams affects the productivity, as changes that require multiply teams to implement need more time (Herbsleb and Mockus, 2003: 486). Furthermore an unstable architecture can lead to confusion of the dispersed teams (Mullick et al., 2006: 208).

Distance-bridging factors for overcoming the technological distance are training or experience with dispersed working situations (Pallot et al., 2010b: 11).

Pallot (2009) states that lack of meaning creates semantic distance (Pallot, 2009). Semantic distance defines the degree to which different terms are related to each other. Distance-bridging factors are classified terms, a taxonomy or ontology. An example would be the semantic web (Pallot et al., 2010b: 10).

The ECOSPACE survey results for the technical dimension show that lack of common description as well as lack of meaning are most significantly impacting collaboration performance. Both factors are necessary for a mutual understanding. On the other hand the survey respondents rated the factor multi-platform setting i.e. different operation system or types of devices as less significantly affecting collaboration performance (Pallot et al., 2010a).

The legal and ethical dimension comprises three different distance types namely ownership, financial and contractual.

Pallot (2009) argues that and Unbalanced Intellectual Property Rights (IPR) approach creates ownership distance (Pallot, 2009). Ownership distance describes the degree to which organizations have different approaches regarding Intellectual Property Rights (IPR). Also local IPR rules, views and opinions differ regarding innovation efficiency. While some argues for open innovation based on open source or creative commons licensing other call for closed innovation and protected IPRs. Distance-creating factors are differing approaches to ownership (Pallot et al., 2010b: 11). When entering a competitive market the unique product features needs to be protected, which can be achieved by trade secrets, copyrights trademarks, mask works or patents (Karolak, 1998: 129–133). Distance-bridging factors are tracking of the individual contributions to innovations e.g. by logging of posts in wikis. Agreeing on common IPR policies also reduces ownership distance (Pallot et al., 2010b: 11).

Pallot (2009) points out that financial distance is created by differing investment regulations (Pallot, 2009). Financial distance depends on the degree to what the involved parties participate and invest in the collaboration. Distance-creating factors are investment policies that create unbalanced investments of the involved parties. Distance-bridging factors are shared risk and common financial agreements of the participating partners (Pallot et al., 2010b: 11).

Pallot (2009) states that contractual distance is created by differing contractual settings (Pallot, 2009). Contractual distance depends on incomplete contracts between the involved partners (Pallot et al., 2010b: 12). The participants’ rights and obligations for all possible circumstances should be specified (Pallot et al., 2010b: 20). Distance-creating factors are incomplete contracts, instable political climate. Distance-bridging factors are formal frameworks for contracts between collaboration partners and experience in international settings (Pallot et al., 2010b: 12).

The AMI@Work survey results for the legal & ethical dimension show, that unbalanced IPR are seen as the most significantly affecting barrier for collaboration performance. Unbalanced IPR means that the participating partners in the collaboration have different approaches for IPR and their objectives regarding IPR differ (Pallot et al., 2010a).

In the next section success factors for GSD are identified.

Success factors

Success factors for GSD can be defined as factors that support overcoming GSD barriers and challenges. Sangwan et al. (2007) introduces five success factors for GSD:

- Reduce ambiguity
- Maximize stability
- Understand dependencies
- Facilitate coordination
- Balance flexibility and rigidity (Sangwan et al., 2007: 10ff)

In collocated settings ambiguity is addressed through informal communication, which in dispersed GSD projects team is hard to accomplish. Due to ambiguity in GSD projects assumptions are made by local teams which later lead to problems, re-planning, redesign and rework. Ambiguities can be found within organizational processes, management practices, requirements or the design. To reduce ambiguity conventions and clearly defined rules for cooperation should be established. Requirements need to be modeled in order to be understood by all members of the dispersed teams. The architecture and components needs to be specified clearly. Task assignments should be clear and when answers are given to certain questions it needs be clarified that the answer is understood (Sangwan et al., 2007: 10–11).

Due to different culture and background, subject matters can be understood differently from remote teams. A solution can be to provide particular training programs or swap staff from one location to another (Sangwan et al., 2007: 11).

Instability causes similar impacts as having ambiguity in the project. In order to maximize stability in GSD projects Sangwan et al. (2007) proposes that the requirements and design phase should be completely finalized and the outcome stable before the development phase is initiated. Especially as change requests take longer than to be resolved than in collocated environments (Sangwan et al., 2007: 12). Also Prikladnicki (2003) sees initial planning as a success factor, which supports the evaluation of distributed projects and the proper assignment of units to projects (Prikladnicki et al., 2003: 274f). It needs to be mentioned that for agile GSD this contradicts to the paradigm of agile development without comprehensive upfront planning.

Further measures to create stability are additional prototypes, UI mock-ups, frameworks and customized development environments (Sangwan et al., 2007: 12).

It is important to understand dependencies in the GSD project. Here, technical aspects like hard requirements that need coordination, because they affect several teams or non-technical aspect like temporal dependencies need to be taken into consideration (Sangwan et al., 2007: 12).

In order to facilitate coordination not only the coordination by communication should be focused but also the coordination through processes, management practices and product line architectures are aspects to consider and elaborate. Prikladnicki (2003) sees the existence of a formal software development process as the most important success factor for distributed projects (Prikladnicki et al., 2003: 274f). There is a trade-off between risk and overhead for example creating a common framework creates efforts, but the framework also eliminates the need to communicate the restraints comprised in the framework (Sangwan et al., 2007: 13).

Flexibility and rigidity need to be balanced, that means on the one hand the practices should be flexible in order to comprise the different backgrounds, development processes, cultures and domain knowledge of the remote teams. On the other hand enough rigidity is needed to assure that aspects like defined processes are followed, requirements are fulfilled and outcome is in line with the agreed architecture. This enables to monitor progress, meet deadlines and ensure quality (Sangwan et al., 2007: 13).

Moreover, the training is a success factor which leads to improved relationships. A further success factor is process engagement in order to get the teams into a first contact. Team integration activities are seen as a success factor, because they improve soft skills, increase trust between the team members and reduce their cultural differences. Communication and feedback was identified as a further success factor, which was facilitated by team integration (Prikladnicki et al., 2003: 274f), (Prikladnicki et al., 2006: 22f).

Carmel (1999) argues that there are five centripetal forces, which stand for success factors for global software team:

- collaborative technology
- development methodology
- product architecture
- team building

Telecommunication infrastructure is said to be the basis for the five forces (Carmel, 1999: 81–85).

In this chapter we have analyzed challenges, barriers and success factors for GSD. We have seen that GSD is a complex field with many dependencies. The problem domain can be divided into the four dimensions, namely structural, social, technical and legal as well as ethnical. As the very essence of GSD is collaboration, we focus on the collaboration barriers that have the biggest impact on collaboration performance. We base this on findings from the ECOSPACE survey published by Pallot (2010). For these barriers adequate success factors for bridging the distance are proposed and are the basis for the further research.

Regarding the barriers affecting collaboration performance, we have identified for the structural dimension lack of clear leadership and lack of incentive as the most important barriers but also unbalanced power, unbalanced expertise and multi-disciplinary settings have a significant impact on collaboration performance. For the social dimension lack of mutual trust is the most important barrier, but also lack of commons, multi-lingual setting need to be taken into consideration to enable efficient collaboration processes in GSD. For the technical dimension lack of common description and lack of meaning are the barriers that most heavily affecting collaboration performance, but lack of media naturalness needs to be reduced in order to support effective collaboration. Last but not least for the technical dimension unbalanced IPR approach has been identified as strongly influencing collaboration performance.

These barriers create distances, based Pallot’s (2009) survey we identified the most important distances, which we will use as an input to our framework.

First configurational distance, factors that bridge the configurational distance like clear leadership and collaboration incentive are important to overcome distance in GSD. These support a shared vision of the dispersed project participants and lead towards working to a common goal and support in avoiding ambiguity.

Second, relational distance – factors that bridge relational distance like team building, team integration facilitate trust, reduce fear and thus raise collaboration performance.

Third, cultural distance – factors that bridge cultural distance and need to be taken into account for successful GSD collaboration is education and training, as well as experience in internationalization. Further distance-compressing factors like online-communities are fostering interaction, thus enhancing mutual understanding and therefore need to be considered when conducting GSD collaboration.

Fourth, conceptional distance – factors that bridge conceptional distance are training and knowledge management e.g. in form of a common taxonomy in a wiki, which enable global teams to create a common meaning and mutual understanding.

Fifth, technological distance is created by the use of different technologies. For successful GSD collaboration common globally accessible and available tools and technologies need to be introduced and established in order minimize technological distance and thus foster collaboration performance.

[...]